Fairfield Today

Anthropic has finally filed its S-1, and buried beneath hundreds of pages of boilerplate legal hedging lies a genuinely remarkable financial picture. The AI lab has raised $141 billion and still holds $131 billion of it, making it, by one estimate, among the most capital-efficient scale-ups in corporate history. A reported $80 billion "loss" turns out to be mostly an accounting artifact from revaluing convertible notes as the company's worth climbed—not cash actually burned. Strip that away, and Anthropic looks profitable, with billions in positive cash flow. But the real story is what the filing reveals about how Anthropic talks about itself versus how it's built. The company has locked in roughly $800 billion in compute commitments to Amazon, Google, Microsoft, Nvidia and others—bets it says it must win or risk falling fatally behind rivals through what it calls "closing the loop," its term for AI systems that help build their own successors. Founders retain effective permanent control through a special voting share, while a Long-Term Benefit Trust meant to safeguard the public interest only gains real power once those founders eventually step back. Meanwhile, warnings about AI's potential for catastrophic harm appear almost entirely in passive voice, tucked between routine risk disclosures about lawsuits and cloud outages. The filing paints a portrait of a company racing to dominate a technology it openly admits could be civilization-altering—while writing about that risk like any other line item.

It started with a loud ultrasound machine, a bag of pig hearts, and a scientist nobody believed. Two decades and millions in federal grants later, that accident became a $4 billion company and an FDA-approved cancer treatment—proof of what patient, unglamorous government funding can build when private investors won't take the risk. But that story is becoming the exception, not the rule. Since President Trump returned to office, nearly $200 billion in federal research funding has been cut, delayed, or redirected across 17 policy sectors and 21 agencies. Grant approvals at the NSF have hit forty-year lows. The NIH has quietly used an AI tool to flag and reject proposals containing words like "racism" or "sexual minority," killing Alzheimer's research simply because it studied why Black Americans face double the risk of dementia. Meanwhile, money is pouring into AI and quantum computing—fields already flush with private capital—while basic science, the unpredictable kind that doesn't promise returns for twenty years, withers. The consequences aren't abstract: an 8-year-old boy with a rare brain tumor lost access to a clinical trial when DOGE cut funding for the consortium running it. He died this month. This piece traces the quiet dismantling of a research system that gave the world GLP-1 drugs, GPS, and stealth technology—and asks what breakthroughs, cures, and companies are being quietly erased from America's future before anyone notices they're gone.